Mobile devices are often expected to perform computational tasks that may be beyond their processing or battery capability. Cloud computing techniques have been proposed as a means to offload a mobile device's computation to more powerful resources. In this thesis, we consider the case where powerful computing resources are made available by utilizing vehicles. These vehicles can be repositioned in real time to receive computational tasks from user-carried devices. They can be either equipped with rugged high-performance computers to provide both computation and communication service, or they can be simple message ferries that facilitate communication with a more powerful computing resource. These scenarios find application in challenged environments and may be used in a military or disaster relief settings. It is further enabled by increasing feasibility of (i) constructing a Mobile High Performance Computer (MHPC) using rugged computer hardware with form factors that can be deployed in vehicles and (ii) Message Ferries (MF) that provide communication service in disruption tolerant networks. By analogy to prior work on message ferries and data mules, one can refer to the use of our first schema, MHPCs, as computational ferrying. After illustrating and motivating the computational ferrying concept, we turn our attention into the challenges facing such a deployment. These include the well known challenges of operating an opportunistic and intermittently connected network using message ferries -- such as devising an efficient mobility plan for MHPCs and developing techniques for proximity awareness. In this thesis, first we propose an architecture for the system components to be deployed on the mobile devices and the MHPCs. We then focus on defining and solving the MHPC movement scheduling problem with sufficient generality to describe a number of plausible deployment scenarios. After thorough examination of the MHPC concepts, we propose a scheme in which MHPCs are downgraded to be simple MFs that instead provide communication to a stationary HPC with powerful computing resources. Similar to the MPHCs, we provide a framework for this problem and then describe heuristics to solve it. We conduct a number of experiments that provide an understanding of how the performance of the system using MHPCs or MFs is affected by various parameters. We also provide a thorough comparison of the system in the dimensions of Computation on the Move and Controlling the Mobility.

<p>Machine learning is a rapidly developing field in computer research. Deep neural network architectures such as Resnet have allowed computers to process unstructured data such as images and videos with an extremely high degree of accuracy while at the same time managing to deliver those results with a reasonably low amount of latency. However, while deep neural networks are capable of achieving very impressive results, they are still very memory and computationally intensive, limiting their use to clusters with significant amounts of resources. This paper examines the possibility of running deep neural networks on mobile hardware, platforms with much more limited memory and computational bandwidth. We first examine the limitations of a mobile platform and what steps have to be taken to overcome those limitations in order to allow a deep neural network to operate on a mobile device with a reasonable level of performance. We then proceed into an examination of ApproxNet, a neural network designed to be run on mobile devices. ApproxNet provides a demonstration of how mobile hardware limits the performance of deep neural networks while also showing that these issues can be to an extent overcome, allowing a neural network to maintain usable levels of latency and accuracy.</p>

<p>In the last decade, people have been surrounded by mobile devices such as smartphones, smartwatches, laptops, smart TVs, tablets, and IoT devices. As sensitive personal information such as photos, messages, contact information, schedules, and bank accounts are all stored on mobile devices today, the security and protection of such personal information are becoming more and more important. Today’s mobile devices are equipped with a variety of embedded sensors such as accelerometer, gyroscope, magnetometer, camera, GPS sensor, acoustic sensors, etc. that produce raw data on location, motion, and the environment around us. Based on these sensor data, we propose novel in-air signature authentication technologies on both smartphone and smartwatch in this dissertation. In-air signature authentication, as an essential behavioral biometric trait, has been adopted for identity verification and user authorization, as well as the development of deep neural networks, has vastly facilitated this field. This dissertation examines two challenging problems. One problem is how to deploy machine learning techniques to authenticate user in-air signatures in more convenient, intuitive, and secure ways by using smartphone and smartwatch in daily settings. Another problem is how to deal with the limited computational resources on today’s mobile devices which restrict to use machine learning models due to the substantial computational costs introduced by millions of parameters. </p> <p>To address the two above problems separately, we conduct the following research works. 1) The first work AirSign leverages both in-built acoustic and motion sensors on today’s smartphone for user authentication by signing signatures in the air without requiring any special hardware. This system actively transmits inaudible acoustic signals from the earpiece speaker, receives echoes back through both in-built microphones to “illuminate” signature and hand geometry, and authenticates users according to the unique features extracted from echoes and motion sensors. 2) The second work DeepWatchSign leverages in-built motion sensors on today’s smartwatch for user in-air signature authentication. The system adopts LSTM-AutoEncoder to generate negative signature data automatically from the enrolled signatures and authenticates each user by the deep neural network model. 3) We close this dissertation with an l0-based sparse group lasso approach called MobilePrune which can compress the deep learning models for both desktop and mobile platforms. This approach adopts group lasso penalty to enforce sparsity at the group level to benefit General Matrix Multiply (GEMM) and optimize the l0 norm in an exact manner. We observe the substantial reduction of compression ratio and computational costs for deep learning models. This method also achieves less response delay and battery consumption on mobile devices.</p>

Šiame darbe analizuojama mobilaus ryšio sistema. Yra tariama, jog tinkle egzistuoja dviejų tipų įrenginiai: mobilūs ir fiksuoti. Fiksuoti įrenginiai yra visados įjungti ir prijungti prie tinklo. Fiksuoti įrenginiai, tai mobiliosios palaikymo stotys (MPS) ir duomenų bazė. Mobilūs įrenginiai (mobilūs kompiuteriai) gauna duomenis iš duomenų bazės, tarpininkaujant palaikymo stotims. Mobilūs kompiuteriai prisijungia ir persijungia prie skirtingų MPS, priklausomai nuo jų skleidžiamo signalo stiprumo. Visi veiksmai atliekami nuosekliai, nepertraukiant galutinio vartotojo darbo. Kuriant tokių sistemų programinius modelius pirmiausia reikia jas išanalizuoti ir suformalizuoti. Viena iš formalizavimo kalbų – PLA, neseniai buvo išplėsta, kad gal���tų aprašyti dinamines sistemas. Šiame darbe supažindinama su naujai išplėsta formalizavimo kalba (dynPLA). Ji naudojama mobilaus ryšio sistemos specifikavimui bei tikrinama ar formalizavimo kalbos išplėtimų užtenka aprašyti minėtąją sistemą. Pasiūlyti dynPLA formalizavimo kalbos patobulinimai bei pateiktas sistemos modelio specifikacijos pavyzdys. / The mobile computing system will be analyzed in this work. It is assumed that network consists of two types of devices: mobile and fixed. Fixed devices are always on and always connected to the network. These devices include, but are not limited to database servers and mobile support stations (MSS). Mobile devices access data from database. They connect and reconnect to different MSS depending on the signal strength. All of this is done seamlessly to the end user and the data transfer is not disturbed. When planning to implement such system in programming language, the first step is analyzing and formalizing it. One of the formalization languages – PLA, was recently extended for the sole purpose of formalization of dynamic systems. This work, “The Usage of dynPLA Formalism for Modeling of Mobile Computing System” covers the usage of new extended formalization language (dynPLA) for mobile computing system, analyzing if dynPLA extensions are sufficient for formalizing the said system. Also some improvements for the dynPLA formalization language will be suggested and the example of the mobile computing system will be given.

L'objectif de cette thèse est de contribuer à l'élaboration d'une méthode de test de systèmes mobiles. L'approche développée est fondée sur la description de tests à l'aide de scénarios et leurs vérifications sur une trace d'exécution. Un scénario modélise le comportement et les interactions que l'on souhaite observer entre un ensemble de nœuds. Les caractéristiques des systèmes mobiles nous ont conduit à représenter un scénario sous deux points de vue différents et complémentaires. Un premier représente des événements de communications entre les nœuds et un second représente la topologie des liens entre ces nœuds. Notre approche est décomposée en deux étapes : une étape de spécification des cas de tests à l'aide de scénarios et une étape de vérification de ces scénarios sur des traces d'exécutions. La première consiste à spécifier à l'aide du langage dédié TERMOS les cas de test de l'application mobile à vérifier. Ce langage TERMOS a été mis en œuvre au sein de l'atelier UML Papyrus. À partir des scénarios décrits de manière graphique, nous générons pour chacun d'eux un automate ainsi qu'une séquence de topologie que nous utilisons dans l'étape suivante. La deuxième étape consiste à vérifier chaque scénario sur des traces d'exécutions provenant de l'application à tester. Pour cela un premier outil recherche les occurrences de la séquence de topologie du scénario dans la trace d'exécution. Pour chacune d'entre elles, l'automate est exécuté et conclut à un verdict. L'analyse de l'ensemble des verdicts d'un scénario permet de détecter les fautes présentes dans le système. / The main objective of this thesis is to contribute to elaborating a mobile system test method. The proposed approach is based on test definition using scenarios and their verification on an execution trace. A scenario modelizes the behavior and the interactions we want to achieve on a set of nodes. Considering the characteristics of mobile systems we represented scenarios from two different but complementary points of view. The first represents communication events between nodes and the second represents the link topology between the nodes. Our approach is composed of two steps : a first step to specify the test cases by using scenarios and a second step to verify these scenarios on execution flows. The first step consists in using the dedicated TERMOS language in order to specify the test cases of the mobile application. The TERMOS language has been developed in the UML Papyrus workshop. Based on the graphically defined scenarios, we generate an automaton for each one of them, as well as a sequence of topologies which we will be using in the next step. The second step consists in verifying each scenario by using execution traces from the application to be tested. Therefore a first tool detects scenario topology sequences in the execution flow. For each one of them the automaton is executed and comes out with a verdict. The analysis of all the verdicts of a scenario allows the detection of faults in the system.